Laplace transforms: the table, the shift rules and initial-value problems
Ten problems on Laplace transforms: e^(3t) and t from the definition with the region of convergence, a sum of a constant, a power and a sine by linearity and the table, the first shift theorem for e^(−2t) cos 3t and t³e^(4t), an inverse transform by partial fractions, an inverse by completing the square into a damped oscillation, a first-order initial-value problem, a second-order initial-value problem with nonzero initial conditions, the transform of a ramp switched on at t = 2, a spring switched on by a unit step at t = π, and the impulse response, step response and final value of y″ + 2y′ + 5y from the transfer function and the convolution theorem, with worked solutions and the mistakes of dropping the initial conditions from the derivative rule, transforming a product as the product of the transforms, shifting the wrong way, transforming u(t − c)f(t) as e^(−cs)F(s) and completing the square in the denominator only.
Before you start
The Laplace transform turns a function of time into a function of a new variable s by an improper integral, and its value is that it turns calculus into algebra. Differentiating y(t) becomes multiplying Y(s) by s and subtracting the initial value, so a linear differential equation with constant coefficients becomes a linear equation for ,Y(s), with the initial conditions already built in. Solve it for ,Y, split the answer into pieces that appear in a short table, and read the solution back off the table. Two shift rules handle exponential factors and forcing that switches on at a later time, and the convolution theorem says that the transform of a system's response is its transfer function times the transform of the input, which is how control engineers and signal processing describe linear systems. These ten problems build the table from the definition, then use it to solve initial-value problems, including ones with switched-on forcing and an impulse.
Definition.,L{f}(s)=F(s)=∫0∞e−stf(t)dt, for the values of s where the integral converges. If ,∣f(t)∣≤Meat, it converges for every .s>a.
Compute L{e3t} and L{t} from the definition, and say for which s each integral converges.
·
Find the transform of .f(t)=4−3t2+5sin2t.
··
Use the first shift theorem to find L{e−2tcos3t} and .L{t3e4t}.
·
Find the inverse transform of
F(s)=(s−1)(s+2)5s+1.
··
Find the inverse transform of
F(s)=s2+4s+132s+3.
··
Solve y′+3y=e−t with y(0)=2 using the Laplace transform.
···
Solve y′′+3y′+2y=4 with y(0)=1 and .y′(0)=2.
··
Find the transform of the ramp that switches on at :t=2:
g(t)={0tt<2t≥2.
···
A spring at rest satisfies ,y′′+4y=u(t−π),,y(0)=0,:y′(0)=0: a constant force switches on at .t=π. Find y(t) and .y(3π/2).
···
A system obeys y′′+2y′+5y=f(t) and starts at rest. Find its transfer function ,H(s), its impulse response h(t) for ,f=δ(t), its step response for ,f=1, and the value the step response settles to.
Answers
L{e3t}=s−31 for s>3 and L{t}=s21 for s>0
,F(s)=s4−s36+s2+410, for s>0
L{e−2tcos3t}=s2+4s+13s+2 for ,s>−2, and L{t3e4t}=(s−4)46 for s>4
f(t)=2et+3e−2t
f(t)=2e−2tcos3t−31e−2tsin3t
y(t)=23e−3t+21e−t
y(t)=2−e−2t
G(s)=s2(1+2s)e−2s
,y(t)=41u(t−π)(1−cos2(t−π))=41u(t−π)(1−cos2t), so y(3π/2)=21
,h(t)=21e−tsin2t, and the step response is ,51−51e−tcos2t−101e−tsin2t, which settles to 51
Worked solutions
Problem 1
Compute L{e3t} and L{t} from the definition, and say for which s each integral converges.
∫0∞e−ste3tdt=∫0∞e−(s−3)tdtCombine the exponentials.
=limb→∞s−31−e−(s−3)bReplace ∞ by ,b, integrate, then take the limit.
L{e3t}=s−31 for s>3The exponential tends to 0 only when ;s−3>0; for s≤3 the integral diverges.
∫0∞te−stdt=[−ste−st]0∞+s1∫0∞e−stdtIntegration by parts with ,u=t,.dv=e−stdt. The bracket is 0 for .s>0.
L{e3t}=s−31 for s>3 and L{t}=s21 for s>0At s=4 these are 1 and .161. The transform of t is the transform of ,1, which is ,s1, divided by s once more.
Problem 2
Find the transform of .f(t)=4−3t2+5sin2t.
L{4}=s4Four times .L{1}.
L{t2}=s32!=s32The table with ;n=2; do not forget the factorial.
L{sin2t}=s2+42The table with :b=2:b on top, b2 below.
F(s)=s4−3⋅s32+5⋅s2+42Linearity: the constants come through unchanged.
,F(s)=s4−s36+s2+410, for s>0At s=1 it is ,4−6+2=0, so .∫0∞e−tf(t)dt=0.
Problem 3
Use the first shift theorem to find L{e−2tcos3t} and .L{t3e4t}.
L{cos3t}=s2+9sThe table with .b=3.
e−2t means ,a=−2, so s becomes s+2F(s−a) with a=−2 is .F(s+2).
L{e−2tcos3t}=(s+2)2+9s+2Replace every ,s, in the numerator too.
,L{t3}=s46, and e4t turns s into s−4.3!=6.
L{e−2tcos3t}=s2+4s+13s+2 for ,s>−2, and L{t3e4t}=(s−4)46 for s>4The shift moves the region of convergence too. At s=1 the first is ;183=61; shifting the other way gives −0.1 (Mistake 3).
Problem 4
Find the inverse transform of
F(s)=(s−1)(s+2)5s+1.
(s−1)(s+2)5s+1=s−1A+s+2BTwo distinct linear factors.
A=1+25+1=2Cover up s−1 and set .s=1.
B=−2−1−10+1=3Cover up s+2 and set .s=−2.
2(s+2)+3(s−1)=5s+1Check by recombining.
f(t)=2et+3e−2tEach s−a1 is .eat. Check: ,f(0)=5, and lims→∞sF(s)=5 as well.
Problem 5
Find the inverse transform of
F(s)=s2+4s+132s+3.
s2+4s+13=(s+2)2+9Complete the square; the roots −2±3i are complex, so there are no real linear factors.
2s+3=2(s+2)−1Rewrite the numerator in terms of s+2 as well.
F(s)=(s+2)2+92(s+2)−31⋅(s+2)2+93A shifted cosine and a shifted sine, with the 3 the sine needs on top.
,(s+2)2+9s+2→e−2tcos3t,(s+2)2+93→e−2tsin3tThe first shift theorem with ,a=−2, read backward.
f(t)=2e−2tcos3t−31e−2tsin3tA damped oscillation: decay rate 2 from the real part of the roots, frequency 3 from the imaginary part. Check: ,f(0)=2, and .f′(0)=−4−1=−5.
Problem 6
Solve y′+3y=e−t with y(0)=2 using the Laplace transform.
sY−2+3Y=s+11Transform both sides; .L{y′}=sY−y(0).
Y=s+32+(s+1)(s+3)1Collect (s+3)Y and divide.
(s+1)(s+3)1=s+11/2−s+31/2Cover-up: 21 at s=−1 and −21 at .s=−3.
Y=s+33/2+s+11/2Combine the two terms in .s+31.
y(t)=23e−3t+21e−tCheck: ,y(0)=2, and .y′+3y=−29e−3t−21e−t+29e−3t+23e−t=e−t.
Problem 7
Solve y′′+3y′+2y=4 with y(0)=1 and .y′(0)=2.
s2Y−s−2+3(sY−1)+2Y=s4L{y′′}=s2Y−sy(0)−y′(0) and .L{y′}=sY−y(0).
(s2+3s+2)Y=s+5+s4Move the initial-condition terms s+2+3 to the right.
Y=s(s+1)(s+2)s2+5s+4Put the right side over s and divide by .(s+1)(s+2).
Y=s(s+1)(s+2)(s+1)(s+4)=s(s+2)s+4The factor s+1 cancels, so the e−t mode never appears for these initial values.
s(s+2)s+4=s2−s+21Cover-up: 2 at s=0 and −22=−1 at .s=−2.
y(t)=2−e−2tCheck: ,y(0)=1,,y′(0)=2, and .y′′+3y′+2y=−4e−2t+6e−2t+4−2e−2t=4. Leaving out the initial conditions gives a solution with y(0)=0 (Mistake 1).
Problem 8
Find the transform of the ramp that switches on at :t=2:
g(t)={0tt<2t≥2.
g(t)=u(t−2)tThe unit step switches t on at .t=2.
L{u(t−2)t}=e−2sL{t+2}The second shift theorem needs the function written in terms of ;t−2; equivalently, shift its argument by .+2.
L{t+2}=s21+s2t+2 is the same ramp, measured from the switch-on time: it starts at the value .2.
G(s)=e−2s(s21+s2)Multiply by .e−2s.
G(s)=s2(1+2s)e−2sCheck from the definition: ∫2∞te−stdt at s=1 is ,3e−2≈0.406, and .G(1)=3e−2. Writing e−2s/s2 drops the jump to 2 (Mistake 4).
Problem 9
A spring at rest satisfies ,y′′+4y=u(t−π),,y(0)=0,:y′(0)=0: a constant force switches on at .t=π. Find y(t) and .y(3π/2).
(s2+4)Y=se−πs,L{u(t−π)}=e−πs/s, and both initial values are .0.
Y=e−πs⋅s(s2+4)1Keep the exponential aside; it only shifts the answer.
s(s2+4)1=41(s1−s2+4s)Partial fractions: 41 at ,s=0, and the rest is s2+4Bs+C with ,B=−41,.C=0.
s(s2+4)1→41(1−cos2t)The table, without the shift.
,y(t)=41u(t−π)(1−cos2(t−π))=41u(t−π)(1−cos2t), so y(3π/2)=21The spring stays at rest until ,t=π, then oscillates between 0 and 21 about its new equilibrium .41. Here cos2(t−π)=cos2t because 2π is a full period; in general the shift must stay inside.
Problem 10
A system obeys y′′+2y′+5y=f(t) and starts at rest. Find its transfer function ,H(s), its impulse response h(t) for ,f=δ(t), its step response for ,f=1, and the value the step response settles to.
,(s2+2s+5)Y=F(s), so H(s)=FY=s2+2s+51At rest, the initial-condition terms are all .0.
h(t)=L−1{(s+1)2+41}=21e−tsin2t,L{δ}=1, so .Y=H. Complete the square and put the needed 2 on top: .21⋅(s+1)2+42.
Step: ,Y=sH(s), which is the transform of ∫0th(τ)dτThe convolution theorem: the response is ,h∗f, and with f=1 the convolution is the running integral of .h.
s(s2+2s+5)1=s1/5−51⋅s2+2s+5s+2A=51 at ;s=0; then B=−51 from the s2 terms and C=−52 from the s terms.
(s+1)2+4s+2=(s+1)2+4s+1+21⋅(s+1)2+42Shift the numerator as well as the denominator.
,h(t)=21e−tsin2t, and the step response is ,51−51e−tcos2t−101e−tsin2t, which settles to 51The final-value theorem agrees: ,lims→0s⋅sH(s)=H(0)=51, allowed because the poles −1±2i are in the left half-plane. The response starts at 0 with slope ,0, as a system at rest must.
Where this goes wrong
1. Dropping the initial conditions from L{y''}
The derivative rule is remembered as "differentiation is multiplication by s", without the terms that carry the starting values.
,y′′+3y′+2y=4,,y(0)=1,y′(0)=2Right so far: Problem 7.
“L{y′′}=s2Y and .L{y′}=sY.”Those are the rules only when .y(0)=y′(0)=0. In general L{y′′}=s2Y−sy(0)−y′(0) and ,L{y′}=sY−y(0), and that is where the initial conditions enter.
,Y=s(s+1)(s+2)4, so y=2−4e−t+2e−2tThis solves the equation, but with y(0)=0 and ;y′(0)=0; it is the right answer to a different problem. The correct solution is .y=2−e−2t. Checking y(0) at the end catches the slip at once.
2. The transform of a product taken as the product of the transforms
Linearity lets sums be transformed term by term, and products look as if they should work the same way.
L{t}=s21 and L{e2t}=s−21Right so far: the table.
“.L{te2t}=L{t}L{e2t}.”There is no product rule. A factor eat is handled by the first shift theorem: .L{te2t}=(s−2)21.
L{te2t}=s2(s−2)1At s=3 this gives 91≈0.111 instead of .1. The product s21⋅s−21 is the transform of the convolution ,t∗e2t=41(e2t−1−2t), not of the product .te2t.
3. Shifting the wrong way
The factor e−2t has a minus sign, and the minus sign is copied into .s.
L{cos3t}=s2+9sRight so far: Problem 3.
“e−2t shifts s to .s−2.”The rule is ,eatf(t)→F(s−a), and here ,a=−2, so s becomes .s−(−2)=s+2. A decaying factor moves the transform to the left, where s is smaller.
L{e−2tcos3t}=(s−2)2+9s−2That is the transform of the growing .e2tcos3t. At s=1 it gives ,−101, while the right value (s+2)2+9s+2 gives ,61, which is what ∫0∞e−te−2tcos3tdt comes to. The wrong formula even has the wrong region of convergence, s>2 instead of .s>−2.
4. u(t − c)f(t) transformed as e^(−cs)F(s)
The second shift theorem is remembered as "a step at c multiplies by e−cs", without the shift of .f.
g(t)=u(t−2)tRight so far: Problem 8.
“.L{u(t−2)t}=e−2sL{t}.”The theorem is :L{u(t−c)f(t−c)}=e−csF(s): the function must be written in terms of t−c first, and .t=(t−2)+2.
G(s)=s2e−2sThis is the transform of ,u(t−2)(t−2), a ramp that starts from 0 at .t=2. At s=1 it gives e−2≈0.135 instead of ;3e−2≈0.406; the missing s2e−2s is the jump of height .2.
5. Completing the square in the denominator only
The denominator is rewritten as ,(s+2)2+9, but the numerator 2s+3 is read as 2s plus a constant.
F(s)=(s+2)2+92s+3Right so far: Problem 5.
“(s+2)2+92s→2e−2tcos3t and .(s+2)2+93→e−2tsin3t.”The shifted cosine has s+2 on top, not .s. The numerator must be written as ,2(s+2)−1, which leaves ,−1, not ,3, for the sine.
f(t)=2e−2tcos3t+e−2tsin3tIts transform is ,s2+4s+132s+7, not .s2+4s+132s+3. Its slope at 0 is ,−1, while the right answer 2e−2tcos3t−31e−2tsin3t has slope .−5.
Print this set: laplace-transforms.pdf (problems, answers, and worked solutions on separate pages).